While more than 50% of all drugs target membrane proteins (MPs), less than 1% of all structures deposited in the protein data bank (PDB) belong to this class of proteins (only about 60 distinct folds). An increasing number of membrane protein structures have been solved in recent years by advanced NMR spectroscopy. However, sparseness of data often limits this promising avenue of structure elucidation. We propose implementation of a novel computational knowledge-based fold assembly algorithm for membrane protein structure elucidation "KAMP". Sparseness of NMR data will be counter-balanced with knowledge-based potentials in a de novo protein structure prediction technique that circumvents the size and complexity limits of earlier methods via initial assembly of disconnected secondary structure elements in the transmembrane region. Our method will facilitate and enable more facile NMR-based structure elucidation of membrane proteins. This will promote understanding of the molecular basis for their function. In turn, development of drugs and therapeutics will be facilitated. E. coli Diacylglycerol kinase (DAGK) will be used as a primary vehicle for developing this combined computational/experimental approach. Aim I: Develop KAMP for de novo Protein Fold Elucidation from Sparse NMR Restraints Aim II: Reconstruct Structure Determination for Membrane Proteins of Known Structure with KAMP for Method Validation and Refinement Aim III: Critically Assess KAMP using Diacylglycerol Kinase as a Rigorous Test Case and Measure Additional Experimental Restraints For this project we have immediate access to considerable bodies of NMR data for diacylglycerol kinase, human glycophorin, outer membrane protein A, and the phospholamban pentamer through established collaborations. The high computational and NMR spectroscopy demands of the proposed research will leverage NIH investments in Vanderbilt University's Advanced Computing Center for Research &Education (ACCRE) and Biomolecular NMR Facility.